Inline formula · one page
The formula t = d / v appears once, with no repeated text. It is an image; the surrounding sentence remains searchable text.
Markdown to PDF
Write, paste or upload Markdown. Choose your PDF settings, then preview and export.
Browser-based conversion; no uploads, login or added watermarks.
Markdown
Load a public Markdown URL
The source must allow browser requests. If it does not, paste the text or upload a local .md file.
Live preview
Review the Markdown structureEverything stays in this browser.
Preview is not the final PDF. Check fonts, page breaks, tables and symbols in your download. See export limits
FROM SOURCE TO DOWNLOAD
How to convert Markdown to PDF
- 01
Add your Markdown
Paste text, upload or drop one .md file. Use a public URL only when its host permits the request.
- 02
Choose the page setup
Set paper and orientation. Open the extra settings for margins, PDF font, size and H2 page breaks.
- 03
Check the actual PDF
Download, open and review the file before sharing. The live preview checks structure, not final pagination.
KNOW WHAT CARRIES OVER
Support and limits
The preview renderer and PDF writer use different paths. The notes below describe the current implementation; they are not a promise that every document converts correctly.
| Markdown content | Live preview | Downloaded PDF |
|---|---|---|
| Headings | Renders Markdown heading levels. | The current PDF writer handles H1–H3 explicitly. H4–H6 are not handled by that heading path; do not rely on them being exported. |
| Paragraphs & emphasis | Renders paragraphs, bold, italic and inline code. | Preserves inline bold, italic and inline code in the PDF. Inline code uses a monospaced font. |
| Lists | Renders ordered and unordered lists. | Writes list markers and text. Nested structure and wrapping need a check in the output. |
| Tables | Displays Markdown rows and columns. | Uses equal-width columns. Rows move to the next page when space runs out; headers do not repeat automatically. Export stops if a row is taller than a page or a formula is wider than its column. |
| Code blocks | Recognised language tags receive syntax highlighting. | Uses monospaced Courier, independently of the selected body font. Indentation and consecutive spaces matched the downloadable sample. Syntax colours are not carried over; check long lines in your output. |
| Links | Shows links in the preview. | Adds clickable PDF link annotations. All five links in the downloadable sample opened their original addresses. |
| Images | Can load Markdown image URLs. | Attempts to embed images. The host must allow cross-origin access; a preview image is not proof that export can embed it. |
| Math & diagrams | Renders supported inline and display math, and Mermaid diagrams. | The checked formulas and flowcharts export as raster images. Those regions are not vector graphics, searchable equations or semantic equation structure for assistive technology; zoom and print sharpness are limited. Ordinary text stays searchable. Read the checked examples and their scope. |
A correct preview does not establish the final PDF layout. An over-wide formula or over-height table row stops export; diagrams may shrink to fit the page, making labels smaller. Changing paper, margins or font does not guarantee a fit. This is a single-document converter, with no batch, folder or vault import.
A FILE YOU CAN CHECK
Test files and results
BASIC ENGLISH DOCUMENT · 20 OCTOBER 2026
The fixed English source includes bold, italic, inline code, an indented code block, a list, a table and five links. The PDF below was downloaded from this page, then opened, rendered and checked against that same source.
WHAT WAS OBSERVED
- Settings: A4 portrait, 24 mm margins, Arial control, 12 pt; H2 page breaks off.
- One PDF page. The title, two list items, two table data rows and final sentence were present. Inline bold, italic and code styles were retained. The code block used Courier; indentation, consecutive spaces and the blank line matched the source.
- All five links, including the list and table links, were clickable and opened their original addresses.
Scope: this one basic English file. It does not test long tables, oversized blocks, formulas, diagrams, images or every Markdown feature. The support table above also includes source-code observations, which are separate from this sample run. Math and diagram results are recorded separately below.
Math and flowchart checks
OUTPUT CHECK · OCTOBER 27, 2026
We downloaded and inspected the PDFs below with this exporter. Default settings were A4 portrait, Arial 12 pt, 24 mm margins and H2 page breaks off. Each result applies to its linked source; it is not a claim that every formula or diagram works.
Simple flowchart · one page
Draft → Review → Publish labels and arrows are visible.
Mixed document · two pages
Inline and display formulas, roots, scripts, table and list formulas, and a branching flowchart are present. Code and the tested dollar amounts stay literal; three distinct link destinations remain annotated.
Longer sources and the settings checked
Long vertical flowchart · one page
All fourteen stage labels and connecting arrows are visible, with the title and closing text on the same page. A long linear chart is not evidence for every complex diagram type.
Long mixed document · five pages
Three formula images, two flowchart images and all twenty-four numbered paragraphs are present. Each page was rendered and inspected.
Other checked combinations: Letter landscape / Georgia 8 pt / 14 mm; Letter portrait / Arial 24 pt / 24 mm with H2 breaks; A4 landscape / Georgia 12 pt / 40 mm with H2 breaks; Letter landscape / Courier 24 pt / 14 mm. These selected combinations do not cover every setting. The mixed source was also imported and downloaded in a 390-pixel browser window; its text, page sizes and embedded image hashes matched the desktop result. This was not a physical-phone test.
When this build stops
Export stops on a reported math or diagram syntax error, a formula wider than its output column, an over-height table row, or a resource limit. The checked mixed source is rejected with A4 portrait, Courier 24 pt and 40 mm margins because a formula cannot fit its table column; it is not silently clipped.
Configured limits: 300,000 input characters, 50 formulas, 12 diagrams, 12,000 characters per diagram and 60 output pages. Visual-size limits are 4 million rendered pixels for one visual and 12 million combined. These guards do not guarantee that every phone has enough memory. Invalid-input, count, image-size, character and page-limit checks were exercised; physical low-memory phones and forced timeout failures were not tested.
ADAPT THE SOURCE, NOT THE WORKFLOW
Examples for different documents
Use the same tool and export steps above. Expand a case only for the preparation that is specific to your source. These source examples are not separate converters or measured output guarantees.
LaTeX mathPrepare formulas inside Markdown
Inline expressions
Use a short expression such as $t = d / v$ within its explanatory sentence. The checked inline sample preserves the formula once as an image, with ordinary text around it.
Display formulas and tables
Use $$…$$ for a separate formula. The mixed sample checks a fraction and root, superscripts, subscripts and formulas inside table cells and a list.
A formula must fit its output column; equal-width table columns can be narrow even on a wide page.
Keep literal dollars in code
Place source code in backticks or a fenced code block. The mixed sample preserves price = "$10" and code containing $x+y$ as literal text. Escape a literal dollar outside code as \$; ambiguous dollar notation has not been exhaustively checked.
Use Markdown math, not a complete .tex project: document classes, bibliographies and custom packages are outside this workflow. Equations export as raster images, not searchable or editable equation objects.
Visual sharpness is limited when zooming or printing. A correct preview does not guarantee a fit on the PDF page; syntax errors, oversized content and resource limits can stop export. Physical low-memory phones remain unverified. See the checked sources, actual PDFs and settings; read the shared export limits.
Mermaid flowchartsKeep labels, arrows and context readable
Choose the direction
The checked graph TD source runs top to bottom. The mixed source checks a left-to-right branch with labelled Yes and No arrows. These are flowchart checks, not coverage of every Mermaid diagram family.
Keep labels readable
A diagram may shrink to stay inside the PDF page. Landscape can help a wide graph, but it does not guarantee that a dense map will be legible.
Read the smallest label at the size your recipient will use.
Keep context beside the chart
Use ordinary headings, paragraphs and lists to explain the chart. The mixed checked source places a flowchart with text, formulas, code and a table in one document.
The checked scope is top-to-bottom and left-to-right flowcharts, a labelled branch and a fourteen-node vertical chain. Other diagram families and rich HTML labels are not established by these tests. Diagrams export as raster images; labels are not searchable or editable diagram text.
Visual sharpness is limited when zooming or printing. A correct preview does not guarantee a fit on the PDF page; syntax errors, oversized content and resource limits can stop export. Physical low-memory phones remain unverified. See the checked sources, actual PDFs and settings; read the shared export limits.
Tables Keep columns readable
Start with a simple table
Use a header row, a separator row and consistent cell counts. Here is the original source example:
| Plan | Seats | Review |
| --- | ---: | --- |
| Starter | 3 | Monthly |
| Team | 12 | Quarterly |
For a wide table, try landscape and inspect each column in the PDF. Shorten headings or split unrelated fields into separate tables before reducing the font. The exporter assigns equal column widths and does not repeat the header on continuation pages.
A readable browser table does not establish that every PDF row fits. Check wrapped text, row boundaries and the association between each value and its heading.
README and documentation Prepare a release handout or runbook
Paste a README or documentation source, or upload one README.md file. Give it a title, ordered instructions and explanations around code and tables. This converter handles one document; it does not collect a repository or maintain a synced documentation website.
Keep commands and structure clear
Use headings, paragraphs and lists for the overview, installation steps and examples. Keep commands in fenced code blocks with blank lines around them. The PDF uses a monospaced code font; preview syntax colors are not exported.
The preview can display more Markdown features than the PDF writer preserves. Follow the support table above for headings, links, footnotes, math, diagrams and other special content.
Check images, badges and local references
Screenshots and badges need image addresses the browser can access. A repository-relative path is not automatically resolved against its original project folder.
Public image hosts can prevent export even when the preview loads. Check images and link destinations in the downloaded PDF. Important diagrams need their own output check, including labels and edges.
Prepare wide and long content
Break a long table into smaller sections when possible. Landscape may help a wide command-output table; portrait is often suitable for a handout. Check every column, code wrap and table break in the actual PDF.
The H2 page-break option can separate major sections. Inspect the result for isolated headings. Choosing orientation, margins or font does not guarantee an oversized block will fit.
Keep the release source
PDF export is a snapshot. Update the README source first and make a fresh PDF for each release or handoff. Keep the source alongside the result so later edits are repeatable.
The tool creates a separate download. It does not write back to GitHub, GitLab or your local repository, and does not synchronize changes with them.
Former guide revision: October 26, 2026. This is preparation guidance, not a separate converter or a measured output guarantee.
Resumes Review contacts, dates and page breaks
Use explicit headings and short bullets
A Markdown source is easy to edit and version; the PDF is a fixed snapshot to send. The retained example below is fictional. Its dates and performance figures are sample text, not results measured by this site.
# Alex Rivera
Software engineer · alex@example.com · [portfolio](https://example.com)
## Experience
### Backend Engineer, Northwind
2021 — Present
- Reduced API latency by 40% by caching hot queries.
- Led the migration from a monolith to three services.
### Junior Developer, Contoso
2019 — 2021
- Maintained the billing service and its test suite.
## Skills
Go, PostgreSQL, Docker, Terraform
## Education
BSc Computer Science, 2019
Use section headings for Experience, Skills and Education, consistent dates, and a filename such as alex-rivera-resume.pdf. Before sending, verify names, dates, contact details and visible link targets in the actual PDF, then check that a heading has not been stranded at the bottom of a page.
This is not an applicant-tracking-system certification. A text layer alone does not establish how a particular recruiting system will interpret the file.
Original resume example last updated: 8 October 2026.
Obsidian notes Clean up vault-specific syntax
Turn the note into ordinary Markdown
Copy or export the note you need. This page does not connect to your vault, sync account, backlinks or graph. The retained example shows syntax to clean up, not an already-converted note:
# Meeting notes
## Decisions
- Ship the beta on the 14th.
- Owner: [[Alex Rivera]] becomes [Alex Rivera](/team/alex).
## Warning
> [!warning] Do not send this draft outside the team.
Replace [[wikilinks]] with plain text or a normal link whose destination you know. Replace embeds with the actual content or an accessible image. Turn a callout marker such as > [!warning] into a heading or bold label if you do not want the marker printed literally.
Review YAML frontmatter, block references and %%comments%% before export. Inline #tags remain text; they do not become vault navigation. Diagrams and formulas still have the output boundaries listed above.
Notion export Prepare an exported Markdown page
Export the page from Notion as Markdown, then use the converter above. This route gives you an editable plain-text source; it does not reproduce the complete workspace or interactive page.
Export one page
Open the Notion page menu (•••), choose Export, then Markdown & CSV. Unzip the download and upload one .md file or paste its text. With subpages included, each non-database page has its own Markdown file; attachments sit in folders and full-page databases export as CSV.
The converter does not connect to a Notion account or import a workspace, folder, ZIP or CSV database. Notion also offers direct PDF export if you want its own page presentation.
Keep paths and links meaningful
Exported file names and inter-page links can contain a page ID. If you rename a file, update its links. Keeping local names does not make those files reachable by a PDF recipient: use accessible web addresses where appropriate.
Local attachment paths are not automatically bundled into this PDF. An image that appears in a preview may still fail at export; check each required image and link.
Replace blocks Markdown cannot represent
Callouts export as HTML; replace a styled box with a blockquote and a bold label, as in the source below. A flattened toggle needs a heading for its summary. Synced blocks can duplicate content across exported pages; inspect the export and retain the intended copy.
Equations may appear as $…$ or $$…$$, and embeds can become plain URLs. These source representations do not guarantee that the PDF preserves an equation or interactive embed.
Check what does not travel
Database CSV files are not Markdown tables. Rebuild the values you need as a Markdown table; database views, filters, sorts and formulas are not restored. Toggles do not fold; comments, backlinks and workspace navigation are not recreated.
Check oversized blocks, page breaks and links in the downloaded file. Shared page settings, support limits and privacy notes apply here too. Public URL loading requests text from that host; it is not a private-workspace connection.
Source example
# Quarterly plan
## Risks
> **Warning** — the vendor contract is still unsigned.
### Timeline
- Draft: 3 October
- Review: 10 October
Former guide revision: October 9 and October 26, 2026. This is preparation guidance, not a separate converter or a measured output guarantee.
ChatGPT conversations Organize selected messages before export
Copy the messages you want to keep and organize them as Markdown. Paste the cleaned text into the editor above, or save and upload one .md file. The tool does not connect to your ChatGPT account or retrieve conversations.
Select, label and summarize
Add a title and speaker labels so exchanges have a clear reading order. Remove repeated prompts, abandoned drafts and irrelevant replies.
For a long thread, start with a short summary of the decisions and actions you kept. Use headings for exchanges, lists for action items and fenced code blocks for commands. Split a very long thread into sections; large tables and code blocks still need an output check.
Review before sharing
Compare dates, names and quoted answers with the original conversation after editing. Independently verify factual claims and citations. Remove private instructions or personal details the recipient does not need.
The PDF is a snapshot, not a replica of the chat interface. Screenshots and chat controls are not imported as conversation structure. Check remote images, wide tables, code wrapping and actual page breaks in the downloaded file.
Source example
# Launch checklist notes
## Question
What should the launch checklist cover?
## Response
- Confirm the release date.
- Review the support page.
- Assign an owner to each task.
Former guide revision: October 26, 2026. This is preparation guidance, not a separate converter or a measured output guarantee.
Local processing and privacy
Local files and remote resources
Pasted Markdown and local .md files are processed in this browser, not uploaded to a conversion server. Loading a URL retrieves content from that address; remote images may contact their hosts. A source can allow preview loading and still block image export.
Page delivery and analytics
Page delivery and analytics have separate network requests. Document contents and conversion results are not included in our analytics events. Read the privacy policy for the page-visit data collected.
Frequently asked questions
Can I upload a PDF or a whole folder?
No. This tool accepts one Markdown document. It does not convert a PDF input or import a folder, batch or Obsidian vault.
Why does a public Markdown URL fail to load?
The URL may be unavailable or its host may block browser requests through cross-origin rules. Paste the text or upload a local .md file if you can access the source yourself.
Does changing the PDF font change the preview?
No. PDF font and page settings apply at export. Review the resulting PDF separately. The font labels map to Helvetica, Times and Courier as explained in the extra settings.
Is Download HTML the same as the PDF?
No. It saves the preview markup as an HTML file. It does not use the PDF writer or guarantee the same appearance when opened elsewhere; referenced images and other resources may still be needed.
What if I need templates, citations or scripted builds?
Those are not features of this interactive exporter. The Pandoc workflow comparison explains the command-line alternative for the same Markdown-to-PDF job.